Melbourne Is Becoming an Evening and Weekend City
Two years of City of Melbourne pedestrian-sensor data show foot traffic recovering unevenly: strongest at the residential fringe, after dark, and around the events calendar, while cafe and restaurant seating still clusters around the old daytime pattern.
The City of Melbourne is not moving like one city anymore.
Across two years of pedestrian-sensor data, the strongest signals are not just about recovery. They are about where and when the recovery is happening. Foot traffic is growing at the residential fringe, the evening period now carries a meaningful share of movement across most precincts, and weekends are doing more work than the old CBD commuter story can explain.
That matters because the city's hospitality footprint still looks heavily shaped by a daytime centre-of-town pattern. The precincts with the thickest weekend crowds are not always the precincts with the most seating relative to demand.
The pattern
The recovery is uneven
On a like-for-like panel of sensors active across both years, citywide foot traffic rose +5.7%. But the gains are concentrated in residential and mixed-use fringe precincts, while the university-anchored areas of Carlton and Parkville, and Docklands, went backwards.
Year-on-year change in foot traffic by precinct
Recent 12 months vs prior 12 months · consistent-sensor panel
Trend is computed only on sensors reporting in both periods (98 of 102 active), so the figure reflects real change in pedestrian volume rather than sensors being added or retired.
Daily rhythm
When the city moves
Averaged across the latest 12 months, the daily rhythm splits cleanly. Weekdays still carry a commuter shape, but the evening band from 5pm to 9pm now accounts for roughly three in ten footfalls in most precincts. Saturday has overtaken Friday as the single busiest day of the week.
Average pedestrian volume by hour of day
Citywide · weekday vs weekend · latest 12 months
Hospitality supply
The mismatch
Comparing 2024 cafe and restaurant seating to weekend foot traffic exposes the tension. The precincts pulling the biggest, growing, after-dark crowds, especially the CBD core, Docklands, and Southbank, offer the fewest seats per weekend pedestrian. Lower bars mean tighter supply relative to demand.
Cafe & restaurant seats per 1,000 weekend pedestrians
2024 seating census / weekend daily footfall · log scale · colour = foot-traffic trend
An intensity proxy, not a utilisation measure. It flags where demand and licensed seating are pulling apart, not whether venues are full.
Event effect
Events change the shape of the city
Against a typical day of about 834,000 footfalls, the event calendar produces swings that dwarf the underlying recovery. The four busiest days in two years are the March Moomba / Grand Prix long weekend. Toggle the markers to see where the calendar punches through the baseline.
Citywide pedestrians per day, across two years
7-day average over raw daily counts · dashed line = typical day
And events have a geography. The Australian Open's footprint is a corridor, not the retail core. The Birrarung Marr riverside path between Flinders Street and Melbourne Park ran far above a normal day across both tournaments.
Melbourne Park corridor: normal day vs Australian Open
Average pedestrians per day · top corridor sensors · Australian Open 2025
The two events most people name first, the Grand Prix at Albert Park and the Melbourne Cup at Flemington, sit in neighbouring councils, so this network sees them only as CBD spillover, never at the venue.
Full comparison
Every precinct, every signal
| Precinct (CLUE small area) | Footfall 12mo | YoY | Evening % | Weekend % | Seats | Seats / 1k wknd |
|---|
Method
Method and caveats
The analysis uses four City of Melbourne open datasets: Pedestrian Counting System hourly counts, Sensor Locations, Cafes & Restaurants with Seating Capacity from the 2024 CLUE census, and the 13 CLUE small-area boundaries. Sensors were joined by point-in-polygon, and event dates were cross-referenced from public listings.
Year-on-year change uses a fixed panel of sensors reporting in both 12-month periods. Event uplift compares an event window's mean daily count to each sensor's own median day.
The limits are worth keeping visible:
- The accessible counts are a rolling two-year window: current trajectory, not recovery versus a pre-pandemic baseline.
- The seating ratio is a supply-tension signal, not occupancy.
- Co-occurring events, especially Moomba and the Grand Prix sharing a weekend, cannot be cleanly separated without a single-event year.